Wavelet-based entropy and complexity to identify cardiac electrical instability in patients post myocardial infarction

Abstract Myocardial infarction (MI) has been long recognized as the main cause of malignant ventricular arrhythmia and/or sudden cardiac death. The region of myocardial scars is related to conduction abnormalities that are rejected as fragmentation of the QRS complexes, which could persist several months after the acute event. In the present work, we evaluated the normalized entropy ( H ) and the statistical complexity ( C ) of QRS complexes, by using the continuous wavelet transform, as an effective method to quantify abnormal alterations in cardiac electrical activity in post-MI patients. We have included the standard 12-leads electrocardiogram (ECG) records of healthy subjects ( C TRL ), n = 48, and MI patients without ventricular tachycardia (VT) and/or fibrillation (VF), grouped in MI healing (MI7), n = 84, and healed (MI60), n = 41, phases. The mean H and C values ( H ‾ and C ‾ ) of each subject were calculated. H ‾ significantly increased and C ‾ significantly decreased (p  C TRL . We integrated all the ECG leads in a single multi-lead criteria ( H ‾ ML and C ‾ ML ). Moreover, we separated MI patients according to the infarcted area in anterior and inferior subsets. H ‾ ML and C ‾ ML showed the same trends as H ‾ and C ‾ for total patients and both infarcted areas subsets, with the advantage that higher values of sensitivity and specificity were obtained. In conclusion, wavelet entropy and statistical complexity applied to ECG records give new insight into the analysis of patients post-MI, who have not suffered VT/VF, in both MI stages, independently of the infarction areas analyzed.